Friend or foe?
16 Oct 2017
AI and machine learning technology is here to stay, but should the financial industry embrace a super-efficient helping hand, or keep it at a safe distance?
![](https://www.assetservicingtimes.com/specialistfeatures/images/MondayOctober1620171508163817.jpg)
In a fast-paced world of ever-expanding technology, the financial services industry has come head on with plenty of challenges, but also with opportunities to innovate.
Through various events, surveys and discussion papers, industry participants have long been discussing a full implementation of artificial intelligence and machine learning technology within their back-office businesses, yet still there is no certain timeframe as to when it will actually happen.
While there are companies that have already started using these new technologies, it is becoming more important for others to jump on board if they’re to remain competitive and innovate.
According to PwC’s 2016 Global Data and Analytics Survey, two thirds of US financial services respondents said they’re not currently ready to rely on machines, because they’re limited by operations, regulations, budgets or resource limitations.
Mark John, head of product and business development for Pershing and broker-dealer services for Europe for BNY Mellon suggests that, so far, there has not been too much disruption, claiming that AI has been more of a “distraction than a real disruption”.
John says: “AI is perceived as a threat to the role of humans in standardised processes, these same processes that are the current core of middle- and back-office functions. As the industry invests into exploratory technology, proof of concepts will become clearer.”
Currently, there is no clear path showing how this new technology will be fully implemented, as it still needs to be developed and challenged. One of the biggest challenges facing the financial services industry, in particular, will be a cultural one.
John comments: “This is a huge shift from the processes that have developed over the years.”
He adds that the crux of the matter lies in “accepting that a decision-making process can be made by a machine, and not a human or collection of humans”.
The biggest disruption would be the removal of the art of debate, replacing it with a single consciousness to decide on the best possible outcomes.”
There has, however, been some positive commentary from companies already using artificial intelligence technologies. Umar Farooq, head of channels, analytics and innovation at J.P. Morgan Treasury Services, suggests that, so far, machine learning is “enhancing current processes”.
Farooq says: “Most machine learning applications in financial services are enhancing current processes and enabling banks to increase growth, drive expense efficiencies and manage risk.”
The word ‘disruption’ tends to suggest a negative impact, but Farooq argues that at J.P. Morgan, disruption is seen as “mostly a positive for banks and clients”.
Another concern around AI often cited in the financial services industry is around not only the role of the human, but the disintermediation of whole sector players. According to Farooq, it is hard to know which players in the industry will become disintermediated because of new technologies.
Again, at this point there little knowledge around how this will play out in financial services. However, for now Farooq explains that middle- and back-office functions have been successful in utilising machine learning and robotics within existing functionalities, rather than replacing them altogether, which points towards a positive future.
He explains: “All types of players, including financial institutions, industry utilities, infrastructure providers, and so on, are actively investing in and investigating new emerging technologies to understand the impact to their business model.”
He adds: “Many of these firms will be able to evolve their business models to take advantage of these new technologies.”
As the future effects of AI development are currently so unknown, as with all technological advancements it is hard to say whether the disruption will ultimately be positive or negative.
John suggests that it is the purpose in which a technology is used that makes it a good thing or not. He notes that governance, and the ways in which AI is applied, will also play a big role in how it affects the industry.
He says: “It is true that repetitive tasks can be replaced by automated functions and history has proven that. The industrial revolutions over the past few hundred years have given way to the rise of factories, to automation, to the digital age.”
“We are now looking to the fourth industrial revolution, where financial process automation will be enhanced by robotics and AI, and underpinned by nano-technology. But, like the factories, our industry will still be managed by people.”
“Controls, quality checks and maintenance are still human tasks. I don’t see that going away anytime in the near future,” he adds.
Although there is a lot of talk and hype around AI and machine learning, how close is the industry to actually utilising these technologies?
Farooq suggests that, in the financial services industry at least, AI is still “several years away from becoming reality”.
However, he explains that big data analytics and machine learning are actively being utilised in a range of areas, from the front office to functions such as fraud screening.
He says: “Automation, or robotics, is also being increasingly used by banks, but it is more prevalent in middle- and back-office functions with manual and repeatable tasks.”
There is currently significant investment going into developing standardisation and automation, as John explains: “The concept goes beyond ‘glorified automation’ as AI seeks to determine certain patterns and makes a logical conclusion from that.”
He says: “When a decision needs to be made, that is when AI comes into the forefront. Analysing a collection of data and historical analytics, as well as looking into causes and resulting effects of a range of past scenarios, can help any fall-out from an already automated process.”
As an example, he suggests that AI could help with identifying repetitive matching errors or fails by applying automated resolution and corrective actions, therefore helping firms comply with regulations such as the Central Securities Depository Regulation.
Some organisations are already seeing success with machine learning, which should give others in the industry confidence around the benefits AI could bring to the table.
The financial services industry needs to ensure the process is not “over complicated” or affecting clients, explains Farooq. He stresses that client experience “has to be at the top banks’ minds at all times”.
“We should not engage in innovation for innovation’s sake,” he says. “It should be a means to an end.”
Rob Palatnick, chief technology architect at the Depository Trust & Clearing Corporation (DTCC), suggests that every financial firm that wants to remain competitive is looking to improve the way it meets its clients’ needs.
Palatnick emphasises that AI is “certainly one of the emerging tools that should be considered, albeit judiciously”.
He explains: “A proper risk-based approach focusing on the client and considering the management, monitoring and support for failover when the automation fails (with manual support available on demand) must be part of the initial design and basic fabric of any intelligence or data-based automation effort.”
Although there is confidence that AI and machine learning will lead to a positive outcome, John suggests the industry should err on the side of caution.
He notes: “The main thing we need to be wary of is the knowledge gap between years of human experience and controls that have been put in place and developed over the years.”
Many of the lessons learnt in the industry have been gained from the “cyclical nature of our business,” he explains.
“We have learnt from past experience in resolving unknown situations and how to apply similar judgements or decisions to help us with future similar situations.”
“A machine can only apply an assessment of what it has learnt like-for-like. Machine learning in that context will still need input from separate experiences and sources to enable analysis and comparison against statistical data.”
In the medium term, John predicts that there will be more of a collaboration of human experience and technology as AI and machine learning mature, rather than an outright replacement of human resources.
However, over a longer period of time, he says: “I would see greater reduction in reliance on middle- and back-office staff, but the pitfall would be placing too much faith too soon in a process that does not have the luxury of being able to challenge from different perspectives such as emotional issues, risk appetite, past experience, and so on.”
Through various events, surveys and discussion papers, industry participants have long been discussing a full implementation of artificial intelligence and machine learning technology within their back-office businesses, yet still there is no certain timeframe as to when it will actually happen.
While there are companies that have already started using these new technologies, it is becoming more important for others to jump on board if they’re to remain competitive and innovate.
According to PwC’s 2016 Global Data and Analytics Survey, two thirds of US financial services respondents said they’re not currently ready to rely on machines, because they’re limited by operations, regulations, budgets or resource limitations.
Mark John, head of product and business development for Pershing and broker-dealer services for Europe for BNY Mellon suggests that, so far, there has not been too much disruption, claiming that AI has been more of a “distraction than a real disruption”.
John says: “AI is perceived as a threat to the role of humans in standardised processes, these same processes that are the current core of middle- and back-office functions. As the industry invests into exploratory technology, proof of concepts will become clearer.”
Currently, there is no clear path showing how this new technology will be fully implemented, as it still needs to be developed and challenged. One of the biggest challenges facing the financial services industry, in particular, will be a cultural one.
John comments: “This is a huge shift from the processes that have developed over the years.”
He adds that the crux of the matter lies in “accepting that a decision-making process can be made by a machine, and not a human or collection of humans”.
The biggest disruption would be the removal of the art of debate, replacing it with a single consciousness to decide on the best possible outcomes.”
There has, however, been some positive commentary from companies already using artificial intelligence technologies. Umar Farooq, head of channels, analytics and innovation at J.P. Morgan Treasury Services, suggests that, so far, machine learning is “enhancing current processes”.
Farooq says: “Most machine learning applications in financial services are enhancing current processes and enabling banks to increase growth, drive expense efficiencies and manage risk.”
The word ‘disruption’ tends to suggest a negative impact, but Farooq argues that at J.P. Morgan, disruption is seen as “mostly a positive for banks and clients”.
Another concern around AI often cited in the financial services industry is around not only the role of the human, but the disintermediation of whole sector players. According to Farooq, it is hard to know which players in the industry will become disintermediated because of new technologies.
Again, at this point there little knowledge around how this will play out in financial services. However, for now Farooq explains that middle- and back-office functions have been successful in utilising machine learning and robotics within existing functionalities, rather than replacing them altogether, which points towards a positive future.
He explains: “All types of players, including financial institutions, industry utilities, infrastructure providers, and so on, are actively investing in and investigating new emerging technologies to understand the impact to their business model.”
He adds: “Many of these firms will be able to evolve their business models to take advantage of these new technologies.”
As the future effects of AI development are currently so unknown, as with all technological advancements it is hard to say whether the disruption will ultimately be positive or negative.
John suggests that it is the purpose in which a technology is used that makes it a good thing or not. He notes that governance, and the ways in which AI is applied, will also play a big role in how it affects the industry.
He says: “It is true that repetitive tasks can be replaced by automated functions and history has proven that. The industrial revolutions over the past few hundred years have given way to the rise of factories, to automation, to the digital age.”
“We are now looking to the fourth industrial revolution, where financial process automation will be enhanced by robotics and AI, and underpinned by nano-technology. But, like the factories, our industry will still be managed by people.”
“Controls, quality checks and maintenance are still human tasks. I don’t see that going away anytime in the near future,” he adds.
Although there is a lot of talk and hype around AI and machine learning, how close is the industry to actually utilising these technologies?
Farooq suggests that, in the financial services industry at least, AI is still “several years away from becoming reality”.
However, he explains that big data analytics and machine learning are actively being utilised in a range of areas, from the front office to functions such as fraud screening.
He says: “Automation, or robotics, is also being increasingly used by banks, but it is more prevalent in middle- and back-office functions with manual and repeatable tasks.”
There is currently significant investment going into developing standardisation and automation, as John explains: “The concept goes beyond ‘glorified automation’ as AI seeks to determine certain patterns and makes a logical conclusion from that.”
He says: “When a decision needs to be made, that is when AI comes into the forefront. Analysing a collection of data and historical analytics, as well as looking into causes and resulting effects of a range of past scenarios, can help any fall-out from an already automated process.”
As an example, he suggests that AI could help with identifying repetitive matching errors or fails by applying automated resolution and corrective actions, therefore helping firms comply with regulations such as the Central Securities Depository Regulation.
Some organisations are already seeing success with machine learning, which should give others in the industry confidence around the benefits AI could bring to the table.
The financial services industry needs to ensure the process is not “over complicated” or affecting clients, explains Farooq. He stresses that client experience “has to be at the top banks’ minds at all times”.
“We should not engage in innovation for innovation’s sake,” he says. “It should be a means to an end.”
Rob Palatnick, chief technology architect at the Depository Trust & Clearing Corporation (DTCC), suggests that every financial firm that wants to remain competitive is looking to improve the way it meets its clients’ needs.
Palatnick emphasises that AI is “certainly one of the emerging tools that should be considered, albeit judiciously”.
He explains: “A proper risk-based approach focusing on the client and considering the management, monitoring and support for failover when the automation fails (with manual support available on demand) must be part of the initial design and basic fabric of any intelligence or data-based automation effort.”
Although there is confidence that AI and machine learning will lead to a positive outcome, John suggests the industry should err on the side of caution.
He notes: “The main thing we need to be wary of is the knowledge gap between years of human experience and controls that have been put in place and developed over the years.”
Many of the lessons learnt in the industry have been gained from the “cyclical nature of our business,” he explains.
“We have learnt from past experience in resolving unknown situations and how to apply similar judgements or decisions to help us with future similar situations.”
“A machine can only apply an assessment of what it has learnt like-for-like. Machine learning in that context will still need input from separate experiences and sources to enable analysis and comparison against statistical data.”
In the medium term, John predicts that there will be more of a collaboration of human experience and technology as AI and machine learning mature, rather than an outright replacement of human resources.
However, over a longer period of time, he says: “I would see greater reduction in reliance on middle- and back-office staff, but the pitfall would be placing too much faith too soon in a process that does not have the luxury of being able to challenge from different perspectives such as emotional issues, risk appetite, past experience, and so on.”
NO FEE, NO RISK
100% ON RETURNS If you invest in only one asset servicing news source this year, make sure it is your free subscription to Asset Servicing Times
100% ON RETURNS If you invest in only one asset servicing news source this year, make sure it is your free subscription to Asset Servicing Times
![Advertisement](../images/bull-252524.png)